Issue Apache hangs every night "server reached MaxRequestWorkers setting"

[abanico]

Server operating system version

CentOS Linux 7.9.2009

Plesk version and microupdate number

Plesk Obsidian 18.0.66, update 2

Hi all!

I'm having an issue since three weeks ago, and it is driving me crazy. I have a Plesk on CentOS 7.9. All working flawlessly for over a year (so flawlessly that I've ever need to reboot the server since installation) and suddenly, every night at 2 or 3 am, Apache stopped working.

The server is not a VPS, nor is it virtualized, it has plenty of RAM (64 GB) and CPU for around 200 domains and it has never passed over 50% of memory usage or 20% CPU usage, so resources are (apparently) the problem.

I have read some threads here about this problem, and I've tried a lot of things, but to no avail:

1. I've tried event and prefork modes of Apache
2. I've raised MaxRequestWorkers (and ServerLimit) on the Apache config, and restarted the server. First to 500, then to 800, and finally to 2000, but this only “delays” the inevitable
3. I've raised the restart interval of Apache in the "Apache configuration" page in Plesk, from "0" (the default setting) to 300, 600 seconds... didn't work either

This is the portion of the Apache error_log from this morning. Instead of fail at 3 pm, I think because of the MaxRequestWorkers, it failed at 06:45 am

[Sun Apr 27 03:42:38.391601 2025] [lbmethod_heartbeat:notice] [pid 24692] AH02282: No slotmem from mod_heartmonitor
[Sun Apr 27 03:42:38.492567 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.492758 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.492943 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.494347 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.498049 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.498549 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.502392 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for webmail.XXXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.504896 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.505218 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.505571 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.505895 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.506215 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.506441 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX_:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.506636 2025] [ssl:warn] [pid 24692] AH01909: RSA certificate configured for default-XXXX:443 does NOT include an ID which matches the server name
[Sun Apr 27 03:42:38.506807 2025] [ssl:warn] [pid 24692] AH02292: Init: Name-based SSL virtual hosts only work for clients with TLS server name indication support (RFC 4366)
[Sun Apr 27 03:42:39.490232 2025] [mpm_prefork:notice] [pid 24692] AH00163: Apache/2.4.6 (CentOS) OpenSSL/1.0.2k-fips mod_fcgid/2.3.9 Phusion_Passenger/6.0.23 configured -- resuming normal operations
[Sun Apr 27 03:42:39.490255 2025] [core:notice] [pid 24692] AH00094: Command line: '/usr/sbin/httpd -D FOREGROUND'
[... OMMITED some lines of modsecurity irrelevant to this problem...]
[Sun Apr 27 06:45:56.424694 2025] [mpm_prefork:error] [pid 24692] AH00161: server reached MaxRequestWorkers setting, consider raising the MaxRequestWorkers setting
[Sun Apr 27 08:12:33.807217 2025] [mpm_prefork:notice] [pid 24692] AH00170: caught SIGWINCH, shutting down gracefully

Also, I don't know if it is relevant, but once I restart the Apache server manually, the cached memory drops dramatically and becomes free memory (see the screenshot)

If anyone can help me with this, maybe to check other things I didn't described here, I will be very grateful, because I'm a little bit lost with this.

Thank you all in advance!

 

[Mikhail_S]

Hello @abanico,

First of all, I would recommend checking what exactly causes so many workers to be spawned. The articles below should help:
Apache service goes down frequently on a Plesk for Linux Server: AH00484: server reached MaxRequestWorkers setting
Apache keeps going down after increasing MaxRequestWorkers many times

If the issue still arises, even when you keep increasing the number of workers, there should be something that causes such behavior. In general, the normal number of workers is calculated like this:

MaxRequestWorkers = (Total RAM - Memory used for Linux, DB, etc.) / average Apache process size

As it is stated here.